Between January 1988 and July 1995, 184 consecutive isolated
subtalar arthrodeses were performed in 174 adults (115 men and fifty-nine
women) who had an average age of forty-three years (range, eighteen
to seventy-nine years) at the time of the procedure. The pathological
findings leading to subtalar arthrodesis included posttraumatic
arthritis following a fracture of the calcaneus (109 feet), a fracture
of the talus (thirteen feet) (Figs. 1-A and 1-B), or a subtalar dislocation (thirteen
feet); primary subtalar arthritis (thirteen feet); and residual
talocalcaneal coalition (eight feet). In addition, twenty-eight
revision subtalar arthrodeses were performed because of the failure
of a previous subtalar arthrodesis. None of the procedures in the current
study were performed because of dysfunction of the posterior tibial
tendon. Our treatment of adult acquired flatfoot secondary to posterior
tibial tendon dysfunction does not include isolated subtalar arthrodesis;
instead, flexible flatfoot deformities are treated with medial displacement
calcaneal osteotomy and either medial soft-tissue reconstruction
or talonavicular arthrodesis, and fixed flatfoot deformities are
treated with triple arthrodesis21.
All patients had had a failure of nonoperative treatment, which
consisted of modification of activities or occupational status,
use of nonsteroidal anti-inflammatory medications, use of an orthosis or
brace, physical therapy, and injection of steroids into the subtalar
joint. When clinical and radiographic evaluation did not confirm
that the pathological findings were isolated to the subtalar joint,
selective injection into the subtalar joint helped to establish
the diagnosis. A sterile mixture of Xylocaine (lidocaine), bupivacaine,
and a corticosteroid was injected into twenty (11 percent) of the
184 feet that subsequently underwent subtalar arthrodesis. The injection
indicated that symptoms were isolated to the subtalar joint and
confirmed that treating the pathological findings in the subtalar
joint with arthrodesis probably would be useful in resolving the
symptoms in the hindfoot.
The average duration of symptoms related to the subtalar joint
was seventeen months (range, four to 126 months). Each patient's
height, weight, activity level, occupation, history of systemic
illness, and smoking status were determined in order to identify
factors influencing the union rate. A standard preoperative clinical
assessment was performed, but without the use of the ankle-hindfoot scale
of the American Orthopaedic Foot and Ankle Society14, which had not yet been published
at the time of the preoperative assessment for most patients. (A preoperative
score according to the scale of the American Orthopaedic Foot and
Ankle Society was assigned retrospectively for patients who had the
arthrodesis before 1994 and prospectively for those who had the
procedure in 1994 or 1995.) All patients had preoperative radiographs
of the foot and ankle.
Operative Technique
All procedures were performed with use of intravenous sedation
and a regional ankle block19 unless
an iliac-crest bone graft was to be used (fifty-four feet), in which
case general anesthesia was used. Each patient received a prophylactic
dose of intravenous antibiotics preoperatively. When necessary (in
fewer than 10 percent [seventeen] of the 184 procedures), an ankle
tourniquet was temporarily applied in order to improve visualization during
the operative approach. The tourniquet was always released at the
time that the subchondral surfaces were assessed for vascularity,
and in no case was it reinflated.
The operative technique for isolated subtalar arthrodesis was
not standardized but instead was tailored to each patient's particular
pathological findings. A lateral approach through a horizontal incision
was used for 150 feet that were treated with in situ arthrodesis,
and a posterolateral approach through a vertical incision was used
for thirty-four feet that were treated with subtalar distraction
arthrodesis. In all patients, the operative procedure involved removal
of all residual cartilage from both surfaces of the subtalar joint
and rigid internal fixation with either one or two screws. Bone
graft was used in 145 feet; the types of graft material included cancellous
autograft (ninety-four feet), structural autograft (twenty-nine
feet), cancellous allograft (seventeen feet), and structural allograft
(five feet). Bone graft was not used in the remaining thirty-nine
feet.
In Situ Subtalar Arthrodesis (Figs. 2-A and 2-B)
In situ subtalar arthrodesis20 is
performed with the patient in the lateral decubitus position. A
straight lateral incision is made over the extensor digitorum brevis
muscle from the anterior aspect of the distal part of the fibula
to the cuboid, in line with the fourth ray. In patients who have
had previous open reduction and internal fixation of a calcaneal
fracture, the same approach to the lateral aspect of the calcaneus
is used again. The extensor digitorum brevis fascia is divided along
its plantar border, immediately dorsal to the peroneal tendons,
and the extensor digitorum brevis is elevated dorsally to expose
the sinus tarsi. The soft tissues, including fat and capsular tissue,
are removed to expose the subtalar joint. When there is evidence
of subfibular impingement (symptomatic compression of the peroneal
tendons between the lateral aspect of the calcaneus and the distal
aspect of the fibula), adequate lateral-wall decompression is performed with
use of an osteotome or chisel.
Next, a lamina spreader is used to improve visualization of the
subtalar joint. An osteotome or chisel is used to remove residual
cartilage and to expose the subchondral bone of the posterior facet of
the calcaneus, the inferior articular surface of the talus, and
the middle and anterior facets of the subtalar joint. Care is taken
to avoid damage to the medial soft-tissue structures. The subchondral
surfaces are denuded adequately to create a vascularized surface.
To increase the surface area of the arthrodesis site, fish-scaling
is performed on the subchondral surface of both the talus and the
calcaneus. The wound is irrigated. If bone-grafting is indicated,
local bone graft (from the site of the lateral-wall decompression
or the anterior process of the calcaneus), iliac-crest graft, or
cancellous allograft is packed into the entire subtalar joint space,
including the spaces between the inferior aspect of the talus and
the middle and anterior facets of the calcaneus.
The subtalar joint is reduced. A partially threaded cannulated
screw is placed from the calcaneus to the talus under fluoroscopic
guidance. A second screw is inserted from the calcaneus into the
talus (again under fluoroscopic guidance) when there is clinical
or radiographic evidence (on intraoperative stress radiographs)
that the first screw has failed to provide adequate stabilization
of the structural graft for bone-block arthrodesis. The wound is
closed in layers; after reapproximation of the extensor digitorum
brevis fascia, the skin is closed.
Bone-Block Distraction
Arthrodesis (Figs. 3-A and 3-B)
Bone-block distraction arthrodesis was used when there was a
decreased talar inclination angle that resulted in anterior ankle
impingement (limited dorsiflexion due to contact between the dorsal
aspect of the talar neck and the anterior aspect of the distal part
of the tibia). With bone-block distraction under the body of the
talus, near-anatomical talar inclination is restored. Structural
autograft or allograft2,4,5,20 was
used in all thirty-four feet that were treated with bone-block distraction
arthrodesis.
The procedure is performed with the patient in either the lateral
decubitus or the prone position. A vertical posterolateral incision
is used, similar to the vertical limb of a standard L-shaped incision for
open reduction and internal fixation of a calcaneal fracture. The
sural nerve is identified, and its main trunk and branches are protected.
The periosteum of the lateral aspect of the calcaneus is elevated
while the peroneal tendons are protected, and a lateral-wall decompression
is then performed with an osteotome.
The subtalar joint is exposed, and a lamina spreader is used
to improve visualization. Rarely, a femoral distractor (Synthes,
Paoli, Pennsylvania) is placed medially, spanning both the ankle
and the subtalar joint, to improve distraction and, thus, exposure.
(The patient typically is placed in the prone position when the
distractor is used.) The joint distractor was used in only nine
(26 percent) of the thirty-four bone-block distraction procedures
performed at our institution. Most of these procedures were performed
in the early part of the series. The decision to use the distractor
was made preoperatively, and thus the patient was in the prone position
from the beginning of the procedure. As our technique improved,
we performed more bone-block distraction procedures with the patient
in the lateral decubitus position and used the lamina spreader exclusively
to gain exposure.
The residual cartilage is denuded from the subtalar joint with
an osteotome, and the surfaces are assessed for adequate vascularity.
The subchondral surfaces are then fish-scaled to increase the surface
area. The lamina spreader is left in place for provisional distraction.
Intraoperative fluoroscopy is used to ensure that adequate distraction
has been achieved. The wound is irrigated. A structural bone graft
is contoured and wedged into the subtalar joint as the lamina spreader
is removed. The size of the structural bone graft is determined
by the amount of distraction required to correct the talar inclination
angle. With the lamina spreader in place, intraoperative radiography
or fluoroscopy is used to document the proper position of the talus. Next,
the posterior height of the subtalar space is measured with a ruler.
The structural graft is fashioned to match this height at its tricortical
base. Intraoperative fluoroscopy is used to confirm that the graft
is in a satisfactory position and that adequate correction has been
achieved.
Under fluoroscopic visualization, a guide-pin is then placed
from the calcaneus through the graft into the talus. The guide-pin
can be inserted to the level of the subtalar joint before insertion
of the graft in order to visually confirm that the screw will penetrate
the center of the graft. The graft is then placed, the guide-pin
is advanced through the graft into the talus, and a fully threaded
cannulated screw is advanced over the guide-pin, across the graft,
and into the talus. Compression is not achieved; instead, the goal
is to restore hindfoot height. The stability of the fixation obtained
with a single screw is assessed intraoperatively. If necessary,
a second screw is inserted with use of the same technique, typically
anterior to the first screw, from the calcaneus into the neck of
the talus. The wound is then closed in layers.
Assessment of Avascularity
Although the present study is retrospective, the senior authors
(L. C. S. and M. S. M.) prospectively maintained a subjective record
of the quality of the bone in the region of the subtalar joint at
the time of subtalar arthrodesis. This independent record served
to identify patients who had clinically important avascularity of
the subchondral bone at the posterior facet of the calcaneus or
the inferior articular surface of the talus. Clinically important avascularity
was arbitrarily defined as at least two millimeters of nonbleeding
subchondral bone noted intraoperatively at the time of subtalar
arthrodesis. The amount of avascular bone was measured from the
level of superficial subchondral bone to the level of bleeding subchondral
bone during preparation of the surfaces for the arthrodesis. Although
radiographs were retrospectively reviewed, the amount of avascular
bone could not be consistently assessed on preoperative radiographs.
Postoperative Management
The procedure was performed on an outpatient basis unless iliac-crest
bone was harvested (fifty-four feet), in which case the patient
was observed overnight for pain control. All patients were discharged to
home wearing a postoperative splint, and all wore a below-the-knee,
prefabricated, removable, fixed-ankle, weight-bearing boot with
a rocker-bottom sole after the wound had healed (typically within
ten to fourteen days after the procedure). Weight-bearing was restricted
for six weeks for patients who had had in situ arthrodesis
and for eight to ten weeks for patients who had had distraction
arthrodesis; thereafter, weight-bearing was permitted as tolerated.
The patients were evaluated both clinically and radiographically
at regular intervals of two, six, and ten weeks and then (if necessary)
at intervals of fourteen, eighteen, twenty-two, twenty-six, and thirty
weeks. Clinical evaluation focused on wound-healing, evidence of
infection, evidence of healing at the arthrodesis site (with stress
applied to the subtalar joint to determine if pain had resolved),
and sural nerve symptoms. Radiographic evaluation consisted of a
lateral radiograph of the foot and one or two Broden subtalar radiographs
to assess healing. Patients with clinical stability and radiographic
evidence of healing (bridging callus or trabeculation) at the union
site were routinely discharged from care at the ten-week interval.
Patients without adequate confirmation of clinical and radiographic
healing were followed at monthly intervals until healing occurred
or delayed union or nonunion was established. Delayed union or nonunion
was indicated by continued clinical symptoms with stress applied
to the former subtalar joint and by lack of bridging callus or trabeculation
at the arthrodesis site as seen on radiographs. In patients who
demonstrated clinical and radiographic evidence of union, the time
to fusion was recorded as the follow-up interval in which union
was determined to have occurred. Although use of the below-the-knee,
prefabricated, removable, fixed-ankle, weight-bearing boot was discontinued
when there was clinical and radiographic evidence of union, the
patient had the option of continuing to use the boot until he or
she felt comfortable enough to wear regular shoes.
Follow-up Assessment
The investigation was approved by the internal review board,
and informed consent was obtained from all patients. Three investigators
who were not directly involved in the operative procedures, including
two of the authors (M. E. E. and H.-J. T.), conducted this retrospective
review. The patients completed a standardized questionnaire at the
time of follow-up to determine (1) the pathological findings that
led to subtalar arthrodesis, (2) the patient's height, weight, smoking
status, and history of systemic illness at the time of subtalar
arthrodesis, and (3) the patient's subjective satisfaction with
the procedure. The questionnaire also was used to identify the patient's
occupational status before the subtalar arthrodesis and at the time
of follow-up.
The clinical evaluation involved a careful history and physical
examination. The history consisted of a review of the questionnaire
with the patient by one of the investigators. The physical examination was
used to assess alignment, pain with stress at the site of the subtalar
arthrodesis, subfibular impingement, the range of motion of the
ankle, pain at the ankle or the transverse tarsal joint, prominent
hardware, and sural nerve symptoms. Finally, a modification of the
ankle-hindfoot scale of the American Orthopaedic Foot and Ankle
Society14 was used to assess functional
outcome. The maximum postoperative score was reduced from 100 to
94 points by elimination of the 6 points assigned to subtalar motion.
The radiographic assessment included an evaluation of lateral
and anteroposterior radiographs of the foot, two Broden radiographs
of the hindfoot, and an anteroposterior radiograph of the ankle.
Radiographic union was determined on the basis of the lateral radiograph
and the two Broden radiographs (as assessed at postoperative visits).
Progressive arthritis of adjacent joints was determined on the basis
of all five views.
Complications were categorized as delayed union or nonunion,
lateral impingement, prominent hardware, sural neuralgia, symptomatic
valgus or varus malalignment of the hindfoot (defined as more than
10 degrees of valgus or more than 5 degrees of varus relative to
anatomical alignment), and infection.
Statistical analysis was performed to determine the impact of
several variables (height, weight, pathological findings leading
to subtalar arthrodesis, smoking status, systemic illness, the failure
of a previous subtalar arthrodesis, the type of bone graft, the
use of one versus two screws for internal fixation, and the presence
of avascular bone at the arthrodesis site) on the time to union
and the rate of union. The methods of analysis included the unpaired
Student t-test, chi-square analysis, one-way analysis of variance
with a Scheffç?²anges test, two-way analysis of variance, and a
forward stepwise multiple-regression analysis. The level of significance
was set at p < 0.05.
The final clinical and radiographic results were determined only
for the 139 patients (148 feet) who were seen at the time of follow-up.
However, data regarding nonunion, infection, and prominent hardware
was available from a review of the charts of all 174 patients (184
feet).
Study Group
Of the 174 patients (184 feet), thirty-one patients (thirty-two
feet) were lost to follow-up: three patients (three feet) died,
four patients (four feet) were contacted but declined to participate,
and twenty-four patients (twenty-five feet) had moved from the area
or could not be located, or both. Of the remaining 143 patients
(152 feet), four patients (four feet) were evaluated with use of
a telephone interview and a follow-up questionnaire without physical
or radiographic examination and 139 patients (148 feet) had a clinical
and radiographic examination at an average of fifty-one months (range,
twenty-four to 130 months) postoperatively. Of these 148 feet, 142
were seen at our institution and six were seen by orthopaedic surgeons
at other institutions closer to the patients' residences. Although
these latter patients did not return to our institution, independent
follow-up (with completion of the standardized questionnaire as
well as a physical examination to determine the ankle-hindfoot score14) was possible with the assistance
of the local orthopaedic surgeons. Radiographs for these patients
were forwarded to our office for review.
Clinical Outcome
In the group of 139 patients (148 feet) who were evaluated both
clinically and radiographically at the time of the present study,
the average score according to the modified system of the American Orthopaedic
Foot and Ankle Society14 (maximum,
94 points) improved from 24 points (range, 0 to 54 points) preoperatively
to 70 points (range, 12 to 94 points) at follow-up. The average pain
score (maximum, 40 points) improved from 2 to 27 points, whereas
the average functional score (maximum, 45 points) improved from
20 to 35 points. Eighty-four percent (117) of the 139 patients had
returned to a level of activity and a work status that were equal
to those before the onset of their hindfoot symptoms, whereas 16
percent (twenty-two) had had no improvement in their level of activity
or work status.
Influence of Pathological Findings on the Results
With the numbers available for study, we found no significant
relationship between the clinical outcome and the pathological findings
that led to subtalar arthrodesis. However, the outcome was most favorable
for patients who had had the procedure because of residual talocalcaneal
coalition and least favorable for those who had had the procedure
because of the failure of a previous subtalar arthrodesis (Table I). Similarly,
we could detect no significant relationship between the rate of
union or the time to union and the pathological findings that led
to subtalar arthrodesis (Table II).
Influence of the Method of
Bone-Grafting on the Results
With the numbers available, we could detect no significant relationship
between the type of bone graft that had been used and the average
ankle-hindfoot score14 (Table III). Similarly,
we could detect no significant relationship between the type of
bone graft and the rate of union or the time to union (Table IV). Although
a nonunion occurred in three of the five feet that had been treated
with structural allograft, the limited number of feet in that group precluded
statistical analysis.
Union
Overall Union Rate
The overall union rate was 84 percent (154 of 184). The union
rate was significantly influenced by smoking, evidence of more than
two millimeters of avascular bone at the subtalar joint, and the failure
of a previous subtalar arthrodesis (p < 0.05 for all). The nonunion
rates that were determined according to clinical and radiographic
criteria differed: thirty feet had clinical evidence of nonunion,
whereas forty-two had radiographic evidence of nonunion. All thirty
feet with clinical evidence of nonunion also had radiographic evidence
of nonunion; thus, twelve feet had radiographic evidence of nonunion
without corresponding clinical evidence.
A forward stepwise multiple-regression analysis was performed
to examine the relationship between nonunion and six potential explanatory variables:
age, gender, type of bone graft, smoking status, avascular necrosis,
and number of screws. The occurrence of nonunion was best predicted with
use of a model using avascular necrosis and type of bone graft (r2 =
0.366, indicating that 36.6 percent of the occurrence was predicted).
None of the other variables contributed significantly to the model.
It should be noted that although the use of fully threaded screws
is generally recommended5, the
use of partially threaded screws is not contraindicated. Although
partially threaded screws were associated with collapse of the graft
in one patient (Figs. 4-A,4-B,4-C, and 4-D), the collapse did not cause a
clinical problem.
Influence of Smoking on the Union Rate
Eighty (46 percent) of the 174 patients were smokers at the time
of the subtalar arthrodesis. Twenty-two (73 percent) of the thirty
nonunions occurred in smokers (Figs. 4-A,4-B,4-C, and 4-D). The union rate for smokers (73
percent; sixty-one of eighty-three feet) was significantly lower than
that for nonsmokers (92 percent; ninety-three of 101 feet) (p < 0.01).
Influence of Avascular Bone
on the Union Rate
A review of intraoperative findings recorded by the senior authors
revealed that seventy-eight (42 percent) of the 184 feet had more
than two millimeters of avascular subchondral bone at the subtalar
joint. All thirty nonunions occurred in this group, resulting in
a 62 percent rate of union (forty-eight of seventy-eight) for feet
that had evidence of avascular bone according to our arbitrary criteria.
Because all nonunions occurred in feet with more than two millimeters
of avascular bone at the subtalar joint, the rate of nonunion was
significantly higher among patients with avascular bone at the site
of the arthrodesis (p < 0.05, chi-square analysis). As anticipated,
the presence of at least two millimeters of avascular bone usually was
observed in association with traumatic etiologies and with the failure
of a previous subtalar arthrodesis. Surprisingly, however, such
bone also was observed in association with nontraumatic etiologies
(Table V).
Influence of Failed Previous Subtalar
Arthrodesis on the Union Rate
The union rate was 86 percent (134 of 156) after primary arthrodesis
and 71 percent (twenty of twenty-eight) after revision arthrodesis;
this difference was significant (p < 0.05). The nonunion rate
was 14 percent (twenty-two of 156) after primary arthrodesis and
29 percent (eight of twenty-eight) after revision arthrodesis. Of
the twenty-eight feet that had a revision arthrodesis, twenty-three
had been referred from other institutions and five had undergone
previous operative procedures at our institution.
Influence of Other Factors
on the Union Rate
Height, weight, the height-to-weight ratio, and systemic illness
did not influence the union rate. Although the number of feet that
had an attempted subtalar arthrodesis adjacent to the site of a
previous ankle arthrodesis was too small for statistical analysis,
our findings suggest that such treatment may be associated with
a diminished rate of union; specifically, two of the six feet that
were so treated went on to nonunion.
With the numbers available for study, we could detect no significant
relationship between the union rate and the number of screws used
for fixation: the union rate was 84 percent (129 of 153) when one
screw had been used and 81 percent (twenty-five of thirty-one) when
two screws had been used.
Analysis of Uncomplicated
Subtalar Arthrodeses
To determine the union rate after subtalar arthrodeses that were
uncomplicated by factors that might diminish the union rate, we
eliminated various subgroups and reanalyzed the data. After eliminating
the feet that had been treated with revision arthrodesis, structural
bone graft (including allograft), and subtalar arthrodesis adjacent
to the site of a previous ankle arthrodesis, the union rate improved
by only 6 percent, from 84 percent (154 of 184) for the overall
group to 90 percent (104 of 116). However, when the patients who
smoked were also eliminated, the union rate improved to 96 percent
(seventy-three of seventy-six).
Radiographic Results
Preoperative and follow-up radiographs were available for 148
(80 percent) of the 184 feet. Preoperative radiographs revealed
that fifty-one (34 percent) of the 148 feet were in patients who
had evidence of degenerative changes in the ankle (twenty-five feet)
or the transverse tarsal joint (twenty-six feet). Follow-up radiographs
revealed that twenty (14 percent) of the 148 feet were in patients
who had evidence of new or progressive degenerative changes in the
ankle (five feet) or the transverse tarsal joint (fifteen feet).
The discrepancy between radiographic and clinical evidence of nonunion
was described earlier.
Complications Other Than Nonunion
Complications other than nonunion included prominent hardware
necessitating screw removal, infection, symptomatic malalignment
of the hindfoot, lateral impingement, and sural neuralgia. Prominent
hardware requiring screw removal was observed in thirty-six (20
percent) of the 184 feet, including thirty (20 percent) of the 150
feet treated with in situ subtalar arthrodesis
and six (18 percent) of the thirty-four feet treated with distraction
arthrodesis. With the numbers that were available for study, we could
detect no significant difference in the prevalence of screw removal
between the feet treated with in situ arthrodesis
(in which compression is achieved) and those treated with distraction
arthrodesis (in which compression is not achieved).
Infection occurred in five (3 percent) of the 184 feet. Two deep
infections were treated with irrigation and d衲idement and intravenous
administration of antibiotics. One of these deep infections resolved,
and the hindfoot went on to union. The second deep infection occurred
in a patient with a history of an open calcaneal fracture. Chronic
osteomyelitis developed despite extensive d衲idements and long-term
antibiotic therapy, and the patient ultimately had a below-the-knee
amputation. The three superficial infections resolved uneventfully
with administration of cephalexin (250 milligrams orally four times
a day for five days) and dressing changes.
Symptomatic malalignment of the hindfoot was observed in nine
(6 percent) of 148 feet. Four feet had excessive varus angulation
(more than 5 degrees of varus relative to anatomical alignment) that
was associated with overloading of the lateral part of the foot,
which remained symptomatic despite orthotic treatment and shoe modifications. Five
feet had excessive valgus angulation (more than 10 degrees of valgus
relative to anatomical ligament) that was associated with clinical
and radiographic evidence of subfibular impingement.
Lateral impingement was noted in fifteen (10 percent) of 148
feet. Eleven feet (including the five feet with symptomatic valgus
malalignment of the hindfoot) had subfibular impingement, and the other
four had impingement at the lateral shoulder of the talus that was
symptomatic during forced dorsiflexion of the ankle or at the end
of the stance phase of gait.
Sural nerve injury was noted in seventeen (9 percent) of the
184 feet. Three of the injuries could be directly related to the
subtalar arthrodesis because they occurred in patients who had not
had previous operative management. All three of these patients had
been managed with distraction arthrodesis through a vertical posterolateral
approach. The other fourteen nerve injuries occurred in patients who
had had previous operative procedures on the hindfoot: twelve of
these injuries occurred in feet that had had open reduction and
internal fixation of a calcaneal fracture either at our institution (three
feet) or at another institution (nine feet), and two occurred in
feet that had had failure of a previous subtalar arthrodesis that
had been performed at another institution.
To the best of our knowledge, our investigation represents the
largest reported series of isolated subtalar arthrodeses in adult
patients. This consecutive series comprised all subtalar arthrodeses
that were performed in a high-volume foot-and-ankle practice during
a seven-year period, regardless of the etiology that prompted the
procedure or the operative technique that was used. Although the
inclusion of a variety of diagnoses and operative methods creates
a heterogeneous study group, it allows for the investigation of
factors influencing the outcome. Furthermore, our study included
the largest reported series of bone-block distraction arthrodeses
and the only reported series of revision subtalar arthrodeses of
which we are aware.
The results of the present study supported our hypothesis that
subtalar arthrodesis is associated with a less favorable outcome
and a higher rate of complications (including nonunion) than has
been previously reported. Traditionally, isolated subtalar arthrodesis
has been associated with high rates of patient satisfaction, low
rates of complications, and an essentially negligible rate of nonunion17,18,24. We have identified several
explanations for our finding that isolated subtalar arthrodesis
has a less favorable outcome than has been suggested in the literature.
First, many of the investigators who have reported the results
of isolated subtalar arthrodesis lacked the objective, accepted
assessment tools now available. Although the ankle-hindfoot scoring system
of the American Orthopaedic Foot and Ankle Society14 combines objective and subjective
criteria, it has been deemed a valid assessment tool and has gained
universal acceptance among most foot and ankle surgeons2-4,10,16,18. With the loss of hindfoot
motion after subtalar arthrodesis, the maximum possible score is
94 points rather than 100 points as originally described14. Mangone et al.16 demonstrated
that this modification of the system represents an accurate and
objective means of determining the outcome of subtalar arthrodesis,
despite the elimination of the 6 points normally assigned to subtalar
motion. Most previous studies on subtalar arthrodesis, like the
current study, have included this modified version of the scoring
system in which the maximum possible score is reduced to 94 points2-4,10,16; one exception was the report
by Mann et al.18, in which the
6 points for subtalar motion were reassigned to transverse tarsal
motion. The average follow-up score in the present study (70 points) was
consistent with those reported in several recent investigations:
Flemister et al.10, in a large
series of isolated subtalar arthrodeses that were performed for
posttraumatic degenerative joint disease, reported an average score
of 74 points; Mangone et al.16,
in a series of subtalar arthrodeses that were performed for a variety
of pathological findings, reported an average score of 77 points;
and Bednarz et al.2 and Burton
et al.4, in series of subtalar
bone-block distraction arthrodeses, reported average scores of 75
and 76 points, respectively.
Second, we investigated all types of subtalar arthrodeses in
this consecutive series of patients, including revision arthrodeses,
arthrodeses performed with use of allograft, and arthrodeses performed
adjacent to the site of a previous ankle arthrodesis. Although these
cases represented only a minority of the procedures in our series,
they appear to have had an influence on the outcome for the entire
group: a nonunion occurred in eight (29 percent) of the twenty-eight
feet that had had a failure of a previous arthrodesis, in three
of the five feet that were treated with structural allograft, and
in two of the six feet in which the subtalar arthrodesis was performed
adjacent to the site of a previous ankle arthrodesis.
Third, smoking has been shown to have a negative influence on
the rate of union7. Eighty (46
percent) of the 174 patients in the present study were smokers,
and twenty-two (73 percent) of the thirty nonunions occurred in
smokers. The rate of union among smokers was significantly lower
than that among nonsmokers (p < 0.01). Bednarz et al.2, in their study of bone-block distraction
arthrodeses, reported a nonunion rate of 14 percent (four of twenty-eight
patients) and noted a strong association between smoking and nonunion.
The rate of nonunion in that study is similar to the overall rate
in our study (16 percent; thirty of 184 feet) as well as the rate
in our subgroup of patients who were managed with bone-block distraction
arthrodesis with use of structural autograft (17 percent; five of
twenty-nine feet).
Fourth, we believe that the presence of avascular bone at the
subtalar joint has a negative influence on the results, particularly
the rate of union. We acknowledge that our assessment of avascular bone
at the subtalar joint is arbitrary. The designation of two millimeters
of avascular subchondral bone is not based on any previously established classification
system, and it is not intended as a proposed classification system;
this arbitrary designation of avascularity simply served as a means to
objectively and quantitatively identify feet with a substantial
amount of avascular subchondral bone at the site of the subtalar
arthrodesis. We recognize that the factors influencing the rate
of fusion at the ankle are not identical to those influencing the
rate of fusion at the subtalar joint; however, in our opinion, the
effect of avascular subchondral bone at the subtalar joint is analogous to
the effect of avascular subchondral bone at the ankle joint as described
by Holt et al.13 and Frey et al.11.
With the numbers available for study, we could detect no significant
relationship between height, weight, the height-to-weight ratio,
or systemic illness and the outcome or the rate of union. Except for
the failure of a previous subtalar arthrodesis, the pathological
findings that prompted the index procedure had an unremarkable influence
on the outcome or the time to union. Aside from the poor results
observed when structural allograft was used, the method of bone-grafting
had no influence on the outcome or the rate of union. The limited
number of structural allografts did not permit statistical analysis;
however, because a nonunion occurred in three of the five feet that
had been treated with a structural allograft, we cannot advocate
the use of structural allograft at this time. With the numbers available,
we could detect no significant relationship between the rate of
union and the number of screws used for fixation: the rate of union
was 84 percent (129 of 153) when one screw had been used and 81
percent (twenty-five of thirty-one) when two screws had been used.
Complications other than nonunion included prominent hardware,
infection, lateral impingement, symptomatic malalignment, and sural
neuralgia. Prominent hardware was not more prevalent among feet
that had been treated with bone-block distraction arthrodesis, and
the rate of screw removal was not different from that associated
with in situ arthrodesis. Despite the lack of intraoperative clinical
and radiographic evidence of hardware prominence, 20 percent (thirty-six)
of the 184 feet required screw removal. An infection occurred in 3
percent (five) of the 184 feet: all three superficial infections
resolved uneventfully, and one deep infection improved after d衲idement
and intravenous administration of antibiotics. The fifth infection,
which developed in a patient who had a history of an open calcaneal
fracture, ultimately necessitated a below-the-knee amputation. Lateral impingement
occurred in 10 percent (fifteen) of 148 feet, either in the subfibular
region or at the lateral shoulder of the talus. Although five of
these cases of impingement were attributable to symptomatic valgus
malalignment of the hindfoot, the other ten occurred despite neutral
alignment and lateral-wall decompression. Symptomatic varus malalignment
was noted in four patients, all of whom had overloading of the lateral
aspect of the foot at the time of follow-up. The malalignment failed
to respond to nonoperative treatment, and the patients were considered
candidates for hindfoot osteotomy. Our results indicate that sural
neuralgia typically is related to previous operative intervention
(such as open reduction and internal fixation of a calcaneal fracture),
but we also noted an increased prevalence among feet that had been treated
with bone-block distraction arthrodesis through a vertical posterolateral
approach, which places the sural nerve at risk.
Mann et al.18 reported that
the prevalence of mild or moderate progression of degenerative changes
in the adjacent joints was 36 percent (twelve of thirty-three) for
the ankle and 41 percent (thirteen of thirty-two) for the transverse
tarsal joint. Those authors reported that the motion of the transverse
tarsal joint was reduced according to objective clinical criteria.
Our radiographic findings support those of Mann et al. In the present
study, the prevalence of progression of degenerative changes in
the adjacent joints was 14 percent (twenty of 148). We surmise that
the prevalence of progression in our study was lower because the
prevalence of degenerative changes on preoperative radiographs was substantial
(34 percent; fifty-one of 148), which makes the determination of
mild radiographic progression less obvious.
It stands to reason that there is an inverse relationship between
the technical complexity of subtalar arthrodesis and the rate of
union. However, even after elimination of what we considered to
be the most complex procedures (those performed after the failure
of a previous subtalar arthrodesis, those performed with use of
structural grafts, and those attempted adjacent to the site of a
previous ankle arthrodesis), the union rate improved by only 6 percent
(from 84 percent [154 of 184] to 90 percent [104 of 116]). When
the group was limited to nonsmokers, the union rate improved to
96 percent (seventy-three of seventy-six), which is consistent with
the findings of several other large studies of isolated subtalar
arthrodesis8,10,18,24. Therefore,
our experience indicates that smoking has a negative influence on
the rate of union following subtalar arthrodesis. Because all thirty nonunions
occurred in feet that had had more than two millimeters of avascular
subchondral bone at the subtalar joint, elimination of that group
would have increased the rate of union to 100 percent (106 of 106).
We therefore believe that it is important to inform the patient
that the intraoperative detection of substantial avascular bone
may diminish the chance for successful fusion. However, the etiology
of the nonunions was multifactorial because 62 percent (forty-eight)
of seventy-eight feet with more than two millimeters of avascular
bone went on to have a successful fusion.
In conclusion, the results of isolated subtalar arthrodesis in
adults are satisfactory, but, based on the recently introduced ankle-hindfoot
scoring system of the American Orthopaedic Foot and Ankle Society14, they are not as favorable as those
reported by previous authors who used a variety of less objective assessment
methods. Factors that had a significant influence on the union rate
included smoking, the presence of avascular subchondral bone at
the arthrodesis site, and the failure of a previous subtalar arthrodesis.
Other factors that probably affect the union rate include the use
of structural allograft and the performance of the arthrodesis adjacent
to the site of a previous ankle arthrodesis. The rate of nonunion
following subtalar arthrodesis probably is determined not by a single
factor but, rather, by a combination of these and other factors.
Note: The authors thank Claude D. Anderson, M.D., for assistance
with data collection; Peter W.-C. Lam, M.D., F.R.A.C.S., for assistance
with the radiographic analysis; and Stuart D. Miller, M.D., for
his contribution to the patient database.